Resource-aware Stream Processing in High Performance Cloud Environment

Today, thousands of ubiquitous computing devices, e.g., mobile devices, wearable devices, and Internet of Things simultaneously generate continuous high-volume data (streaming data). To analyze these data in real-time, high performance cloud environment has emerged. It provides tremendous and flexible computing power by provisioning and scheduling computing resources autonomously. This work focuses on resource scheduling issues among parallel high-volume streaming applications. The robustness and fairness of resource-aware stream processing in high performance cloud environment are analyzed. As the rates of the streams vary, we propose a resource scheduling strategy that is capable of dynamically allocating resources. The strategy is shown to be robust about a system optimum characterized by a proportional fairness criterion.